The invention concerns substantially spherical particles of lithium silicates wherein in particular the mechanical properties are significantly improved by means of special additives (up to 5 wt. % tellurium and/or tellurium compounds), without unfavorably influencing the property of the particle material of releasing tritium upon neutron bombardment, so that said particles can advantageously be used as breeder material for tritium.
Lithium silicates present one option for direct production in the reactor of the tritium needed in nuclear fusion. The ##EQU1## Li present in the lithium silicate thereby traps the neutrons generated in the fusion process and transforms itself into helium and tritium in a nuclear reaction (breeeding reaction): ##EQU2## The tritium generated is then used in the reactor in a plasma in the nuclear fusion reaction, whereby it releases neutrons which can then be used for the breeding reaction: ##EQU3## The ##EQU4## Li needed for the breeding reaction can be provided in a number of forms, for example as lithium silicate, whereby lithium orthosilicate is preferred (viz. Fusion Technology, 1990, page 978).
In the case of solid lithium carrier agents there are a number of possibilities of arranging these in a packet bed (viz. E. Proust et al., 2nd Int. Symp. on Fus. Technol., Karlsruhe, Jun. 2-7, 1991), whereby the arrangement in the form of a pebble bed presents only one option.
If the option of a pebble bed is desired as the arrangement for a fusion reactor, the particles can be directly produced from the melt according to a known process (spray process, centrifuging process and the like) in a size of up to some 1000 .mu.m.
In production processes whereby spherical particles are directly produced from the melt, micro-cracks and cavities are normally observed which may affect up to app. 10-20% of the particles produced. These particles are classed as defective, since they only possess little mechanical capacitance. For example, the maximum pressure capacitance of such particles of lithium orthosilicate particles falls from 8 Newton to 2 Newton, when the above-described defects occur (viz. J. Nuc. Mater., 155-57 (1988), 451). Since the defective particles are statistically distributed among the faultless particles, the totality of all particles possesses a considerable variation in mechanical properties. The pressure test described below serves to assess mechanical capacitance.
If the particles are subjected to constantly changing temperature and pressure strain during use, a part thereof can be destroyed and their suitability thus be called into question. This case is given, if the particles are used as breeding material for tritium in a fusion reactor. Their suitability as breeding material is determined by the thermocycling test described below.
Simulation tests for the temperature and pressure capacitance of a pebble bed in the blanket of a fusion reactor conducted with known spherical lithium silicates show that up to 10% of the particles used do not resist the strain and break (viz. Fusion Technology, 1990, p. 822). Particularly the above-described defective particles are thus affected.
Tests with additions of silicon to the lithium orthosilicate and/or Compounds thereof as well as aluminium and/or aluminium compounds were already conducted at laboratory level a number of times; they resulted in improvements, but these could only be obtained upon subsequent treatment under precisely defined conditions (viz. Fusion Technology, 1990, p.822). The subsequent treatment consists of heating the pebbles of Li-orthosilicate in a rotating tube furnace to a temperature of precisely 1030.degree. C. and then swiftly cooling them again to below 1024.degree. C. At temperatures above 1024.degree. C. the time during which the Si-rich phase is liquified (viz. FIG. 5) may thereby not exceed 5 minutes, since the pebbles would otherwise fuse with each other. Pebbles which do not reach the critical conditions so defined do not achieve an improvement of their mechanical properties. The realization of these precisely defined conditions on a large scale is therefore extremely difficult.